During the course of its development a sea urchin embryo constructs a pair of calcareous endoskeletal spicules. These spicules are rod shaped, mineralized structures which are calcitic assemblages of calcium carbonate (95%) and magnesium carbonate (5%) with an occluded proteinaceous integral matrix. The spicules are synthesized by a well characterized single tissue type, the primary mesenchyme cells. The calcite of each spicule rod is aligned along a single crystal axis, appearing as if each spicule is composed of one crystal of calcite. However, the spicule has greater flexural strength than a single crystal of calcite and it fractures as if it is made up of many microcrystals(1, 2, 3) . Persuasive biophysical evidence indicates that the proteins embedded within the mineral phase of the spicules, the integral spicule matrix proteins, cause these interesting physical characteristics(1, 4, 5, 6, 7) . It is believed that these proteins interact with specific faces of the calcite crystal when occluded within the mineral, and it is through these interactions that control of spicule growth occurs. However, the precise molecular mechanisms underlying interactions with noncollagenous integral matrix proteins that control the formation and the physical properties of mineralized tissues remain unknown. A question basic to our understanding of these mechanisms in mineralizing tissues is what is the nature of the noncollagenous integral matrix proteins that are intimately associated with the mineral portion of these tissues.

Many proteins have been identified and characterized as noncollagenous integral matrix proteins of hard tissues from various vertebrate and invertebrate organisms (most numerously from vertebrate bone and teeth). These types of proteins usually share the general properties of being soluble and acidic(8, 9) . However, it has proven difficult to ascribe to these types of proteins precise functions within the cell that synthesize them. In addition, while many noncollagenous integral matrix proteins have been identified, it has also proven difficult to determine with much certainty what particular integral matrix proteins are contained within a given mineral phase at a given stage of development. Many of these difficulties are due to the complex structures and dynamics of the mineralized tissues most widely studied, i.e. vertebrate bone and teeth.

Sea urchin spicule formation, on the other hand, is particularly well suited to ask these sorts of basic questions. The spicules are synthesized by a single well characterized tissue type and they are relatively simple mineralized structures that do not have the complex dynamics of vertebrate bones or teeth. The sea urchin embryo is also very amenable to biochemical and molecular experimental analysis. Much is known about the cell and developmental biology of sea urchin embryos and particularly about the differentiation of the cell lineage which synthesizes the spicules (for some reviews, see (10, 11, 12, 13, 14) ). Benson et al.(15) reported that the spicule matrix within the mineralized spicules is arranged in concentric sleeves of irregular fibrillar proteinaceous material with some interconnections between the layers of matrix material. The concentric layers of the spicule matrix are also reflected in the concentric layered architecture of the mineralized sea urchin spicules(2) . Benson et al.(16) and Venkatesan and Simpson (17) also identified 8-10 different proteins as comprising the integral spicule matrix. In these studies, one-dimensional SDS-PAGE ()was used to resolve the spicule matrix proteins. In addition, both reports demonstrated that most of these detected proteins are N-linked glycoproteins. Total amino acid analysis revealed Strongylocentrotus purpuratus spicule matrix proteins are rich in acidic amino acids(16) . This amino acid composition is similar to that of other integral matrix proteins closely associated with the mineral portion of other mineralized tissues(8, 9) .

Two different cDNAs that encode two different spicule matrix proteins have also been isolated from S. purpuratus cDNA expression libraries. The first cDNA isolated encodes a protein designated SM50 which has a deduced amino acid sequence with a molecular mass of approximately 50 kDa(18, 19, 20, 21) . The second cDNA cloned encodes a protein designated SM30 which has a derived amino acid sequence with a molecular mass of approximately 30 kDa(22) . The predicted chemical characteristics of the deduced amino acid sequences of SM50 and SM30 cDNAs are somewhat different. The cloned SM50 cDNA encodes a protein with an alkaline pI without any consensus N-linked glycosylation site(21) . The cloned SM30 cDNA encodes an acidic protein that contains a consensus N-linked glycosylation site(22) . Both SM50 and SM30 transcripts are expressed exclusively in the primary mesenchyme cells(19, 22) . Sucov et al.(18) have shown that SM50 is a single copy gene in S. purpuratus. Alternatively, there is experimental evidence that there is a small family of SM30 protein genes. Akasaka et al.(23) presented Southern blotting analysis indicating that there are between two and four copies of SM30 genes present in the S. purpuratus haploid genome. Akasaka et al.(23) further demonstrated that an isolated S. purpuratus genomic clone contains two different SM30 genes that are arranged tandemly. These two SM30 genes were designated SM30- and SM30-. Initial characterization of the genomic regulatory regions of the SM50 gene and the SM30- gene have also been done(23, 24, 25, 26) .

In addition to these two genes that have been shown directly to encode two spicule matrix proteins, a recent report by Harkey et al.(27) characterizes a gene encoding a nonglycosylated 27-kDa protein, designated PM27, that is closely associated with growing sea urchin spicules. While they did not show directly that the PM27 protein is an integral spicule matrix protein, they do show PM27 expression and biochemistry are similar to what one might expect from a spicule matrix protein. In addition they show that PM27 has some sequence similarity to SM50; Harkey et al.(27) point out that portions of PM27, SM50, and the Lytechinus pictus homologue of SM50 (designated LSM34 by Livingston et al.(28) ) also have some similarity to the carbohydrate recognition domain (CRD) of a number of C-type lectin proteins.

The present paper enumerates more accurately the complexity of the S. purpuratus spicule matrix proteins and more fully characterizes these proteins. These studies provide a biochemical foundation important for the study of the noncollagenous integral matrices of mineralized tissues. Our findings also complement the previously mentioned biophysical studies that examined occluded matrix proteins of sea urchin embryonic and adult mineralized structures and their roles in regulating mineralized tissue formation and structure (1, 4, 5, 6, 7) . The studies in the present paper reveal that there are 12 spicule matrix proteins that radiolabel intensely with [S]methionine and approximately three dozen other spicule matrix proteins that are less highly radiolabeled. The majority of the spicule matrix proteins have an acidic pI, while several other moderately radiolabeled to less radiolabeled spicule matrix proteins have a more alkaline pI. Polyclonal antisera that react specifically with the proteins encoded by the previously cloned SM50 and SM30 spicule matrix cDNAs were generated. Western blotting analysis using these antisera identify the SM50 and SM30 proteins. In addition, comparisons are made between the embryonic spicule matrix proteins and the adult spine integral matrix proteins. Further analysis of the protein encoded by SM30- reveals that a portion of the SM30 proteins is similar to the CRD of the C-type lectin family of proteins.

EXPERIMENTAL PROCEDURES

Culturing of Sea Urchin Embryo

Isolation of Spicule Matrix and Spine Matrix Protein

Radiolabeled S. purpuratus spicule matrix protein was isolated from micromeres cultured in seawater with 4% horse serum and [S]methionine. The isolation and culture of micromeres was done essentially as described by Benson et al.(30) . The micromeres isolated from about 2 10 16-cell embryos were cultured in four 100-mm Petri plates, each containing 10 ml of seawater containing 4% horse serum that had been dialyzed against seawater. Two hundred µCi of [S]methionine (1000 Ci/mmol; Amersham) were added to each plate just prior to the onset of spiculogenesis and left in the medium until the time of harvest (24-72 h). At the conclusion of labeling, carrier spicules from whole embryos were added to the cultures, and the adherent spicules of the culture were scraped from the Petri plates. The spicules were washed with and then placed into 3.5% sodium hypochlorite overnight at room temperature. They were then washed with 5-7 changes of dHO. After the final wash, the spicules were suspended in 1 ml of dHO. An aliquot was removed to quantitate the amount of radioactivity incorporated into spicule matrix. To prepare the radiolabeled spicule matrix protein for each two-dimensional gel, 5.0 10 dpm of the radiolabeled spicule sample was trichloroacetic acid precipitated with 10 µg of cytochrome c carrier; this procedure dissolves the calcite and precipitates the spicule matrix protein. The pellet was then washed with acetone to remove residual trichloroacetic acid. The pellet was dried and dissolved in the appropriate sample buffer.

Isolated spicule matrix proteins were labeled in vitro with biotin following the protocol described by Meier et al.(31) using the biotinylation agent NHS-CC-biotin purchased from Pierce. The labeled proteins were localized using an enhanced chemiluminescence protocol described by Nesbitt and Horton (32) with the exception that the blocking agent used was 0.1% fish gelatin (Amersham), 0.8% bovine serum albumin, 0.02% Tween 80, 10 mM Tris, pH 8.0, 100 mM NaCl, and the dilution of strepavidin-horseradish peroxidase (Amersham) used was 1:3000.

Glycosidase treatment of spicule matrix protein was carried out at 37 °C overnight using endoglycosidase F/N-glycosidase F purchased from Boehringer Mannheim following the protocol provided by the manufacturer.

Mild alkaline hydolysis -elimination of O-linked carbohydrate moieties on spicule matrix proteins was done essentially as described by Florman and Wassarman(33) . Spicule matrix proteins were incubated in 5 mM NaOH for 24 h at 37 °C, neutralized with HCl, and then analyzed by Western blotting analysis as described below. Serine and threonine O-glycosidic linkages are known to be labile in alkaline conditions(34) .

Two-dimensional Gel Electrophoretic Separation of Spicule Matrix Proteins

Figure 2: Two-dimensional gel electrophoresis of radiolabeled sea urchin spicule matrix proteins. [S]Methionine radiolabeled spicule matrix proteins were separated by two-dimensional gel electrophoresis. The first dimension separates the proteins on the basis of their pI using different conditions for each of the three gels. The second dimension of all three gels is a 10% SDS-PAGE gel. A, the first dimension of this gel was a nonequilibrium pH gradient tube gel using ampholytes with a pH range of 2.5 to 6.5. The gel was run in the acidic direction for 20 min before placing it on the second dimension. B, the first dimension of this gel is an isoelectric focusing gel using ampholytes with a pH range of 2.5 to 6.5. The gel was run in the acidic direction until the proteins came to equilibrium at their pI (about 2 h), before placing it on the second dimension. C, the first dimension of this gel was a nonequilibrium pH gradient gel using ampholytes with a pH range of 3 to 10. The gel was run in the basic direction for 20 min before placing it on the second dimension. After all the gels were run in both dimensions, the gels were processed for fluorography and exposed to x-ray film. I, II, and III labels indicate the three spicule matrix proteins that were used to align the proteins that were common among the gels in A-C. NEPHGE, non-equilibrium pH gradient gel electrophoresis.

Figure 6: Two-dimensional Western blots of spicule matrix proteins using the anti-SM30 and anti-SM50 antiserum. A, 0.5 µg of unlabeled spicule matrix protein was separated in the first dimension on an isoelectric focusing gel using ampholytes with a pH range of 2.5 to 6.5 as described under ``Experimental Procedures.'' The second dimension was a 10% SDS-PAGE gel. The proteins were subjected to Western blotting using the anti-SM30 antiserum. B, 0.5 µg of unlabeled spicule matrix protein was separated in the first dimension on a nonequilibrium pH gradient gel run in the basic direction. The pH range of ampholytes used was 3 to 10. The second dimension was a 10% SDS-PAGE gel. The proteins were subjected to Western blotting analysis using the anti-SM50 antiserum. The immunoreactive proteins in these two blots were visualized using chemiluminescence. NEPHGE, non-equilibrium pH gradient gel electrophoresis.

Generation of Polyclonal Antisera

A rabbit polyclonal antiserum raised against all of the spicule matrix proteins was generated following the procedure described by Benson et al.(16) and using spicule matrix protein isolated from embryonic spicules as immunogen.

Western blotting of one- and two-dimensional gels was done as described by Towbin et al.(40) . Chemiluminescent detection of immunoreactive proteins was done following the directions of the manufacturer (Amersham) except that 0.1% fish gelatin (Amersham), 0.8% bovine serum albumin, 0.02% Tween 80, 10 mM Tris, pH 8.0, 100 mM NaCl was used as the blocking solution. The anti-SM30 antiserum was used at a 1:2000 dilution and the anti-SM50 antiserum was used at a 1:1000 dilution for the chemiluminesence blots.

Detection of immunoreactive proteins using alkaline phosphatase-conjugated secondary antibody and 5-bromo-4-chloro-3-indolyl phosphate/nitro blue tetrazolium chloride (BCIP/NBT) as substrate was done as described by Richardson et al.(29) . The anti-total spicule matrix antiserum was used at a 1:1000 dilution, the anti-SM30 antiserum was used at a 1:250 dilution, and the anti-SM50 antiserum was used at a 1:100 dilution for these blots.

In Vitro Translation of SM30 RNA by Xenopus Oocytes and Reticulocyte Lysate

The microinjection of the synthesized RNA into Xenopus oocytes and the collection of [S]methionine-labeled secreted proteins was done as described in Livingston et al.(28) . The immunoprecipitation of [S]methionine-labeled proteins using the anti-SM30 antiserum or its preimmune serum at a dilution of 1:100 was also done as described by Livingston et al.(28) . SM30 RNA was translated by reticulocyte lysate using the Ambion Reticulocyte Lysate kit following the protocol provided by Ambion.

RESULTS

Enumeration of the Spicule Matrix Proteins

We decided, alternatively, that radiolabeling the proteins with [S]methionine was a better way to visualize these proteins. We found culturing isolated micromeres in the presence of [S]methionine to be the most effective method of labeling sea urchin spicule matrix proteins. Harkey and Whiteley (43) demonstrated that the time course of spicule formation in S. purpuratus micromere cultures closely matches the time course of events in whole embryos and that the two-dimensional gel pattern of proteins synthesized by cells from micromere cultures closely matched that of primary mesenchyme cells in intact sea urchin embryos.

The radioactively labeled spicule matrix proteins synthesized in culture were isolated and then fractionated by gel electrophoresis. A typical one-dimensional gel fractionation of these radiolabeled proteins is displayed in Fig. 1A. This pattern is very similar to the pattern we observed for spicule matrix proteins isolated from whole embryo spicules that were labeled in vitro with biotin. The pattern of biotinylated proteins is displayed in Fig. 1B. The most noticeable differences in the radiolabeled and biotinylated spicule matrix patterns is that a band at approximately 27 kDa is more prevalent on the radiolabeled protein gel and there is strongly labeled material at 200 kDa on the biotinylated protein gel. The one-dimensional SDS-PAGE patterns observed here are very similar to the patterns shown by one-dimensional SDS-PAGE silver staining and radiolabeling spicule matrix preparations of whole embryos that were observed by Benson et al.(16) and Venkatesan and Simpson(17) . These previously published one-dimensional SDS-PAGE patterns did not have the prominent bands at 200 kDa that is present on the biotinylated spicule matrix SDS-PAGE pattern. Taken together these results indicate that radiolabeling spicule matrix proteins from micromere cultures generates a labeling pattern similar to patterns obtained with stains. Furthermore, given its greater sensitivity and clarity, we found radioactive labeling the spicule matrix proteins better for further two-dimensional gel analysis.

Figure 1: One-dimensional SDS-PAGE of radiolabeled and biotinylated spicule matrix protein. A, [S]methionine radiolabeled spicule matrix proteins were separated by a 10% SDS-PAGE gel. The radiolabeled proteins were then localized using fluorography. B, biotinylated spicule matrix proteins were separated on a 10% SDS-PAGE gel and then electroblotted to nitrocellulose membrane. The biotinylated proteins were then localized on the membrane using an enhanced chemiluminescence protocol described under ``Experimental Procedures.''

The results of two-dimensional gel analysis of the radiolabeled spicule matrix proteins are shown in Fig. 2. In Fig. 2A, the spicule matrix proteins are separated in the first dimension on the basis of their pI using a non-equilibrium pH gradient gel in which the proteins migrate in the acidic direction for a short time (about 20 min). The pH range of the ampholytes used in this gel was between 2.5 and 6.5. In Fig. 2B, the spicule matrix proteins were again separated in the first dimension on the basis of their pI, using the same pH range of ampholytes as in Fig. 2A. However, in this gel, the proteins were allowed to migrate to their equilibrium pI. In Fig. 2C, the proteins were also separated on the basis of their pI in the first dimension. The pH range of the ampholytes used here, however, was 3 to 10, and the first dimension was run in the opposite, basic direction for only short time (about 20 min), which does not allow the proteins to migrate to their equilibrium pI. If these proteins were allowed to migrate to their equilibrium pI, the most alkaline proteins would migrate off the alkaline end of the gel.

It is apparent most of the proteins are acidic, with molecular masses ranging from about 20 to over 100 kDa. The 8-10 bands previously identified by Benson et al.(16) and Venkatesan and Simpson (17) as S. purpuratus spicule matrix proteins have apparent molecular masses similar to many of the proteins separated by the two-dimensional gels presented here. All of the previously identified bands can be accounted for by proteins fractionated on the gels presented in Fig. 2. Fig. 2, A and B, show that the majority of spicule matrix proteins, as well as all of the most highly radiolabeled proteins, are indeed acidic. Fig. 2A reveals a bit more complexity of the matrix proteins than seen in Fig. 2B, especially for the proteins migrating between apparent molecular masses of 30 and 69 kDa. This result is probably because the most heavily radiolabeled proteins on Fig. 2A do not streak out as much before they reach their isoelectric point as they do in Fig. 2B. Since most of the spicule matrix proteins are glycoproteins, one expects to see streaking on equilibrium gels(41) . This streaking presumably occurs because of the heterogeneity in the charge provided by the carbohydrate moiety. Since the first dimension of the gel in Fig. 2B has been run to equilibrium, it provides a more accurate relative comparison of the pI values of the various spicule matrix proteins.

The gel shown in Fig. 2C resolves spicule matrix proteins that have a more alkaline pI. There are a number of spicule matrix proteins in the more acidic portion of the gel in Fig. 2C that also resolved in the more alkaline portion of the gels in Fig. 2, A and B. The proteins labeled I, II, and III ( Fig. 2and Fig. 3) were used as landmarks to orient the proteins in all three of the gels. Fig. 2C demonstrates that there are several spicule matrix proteins, with medium to low radiolabeling, which have a somewhat more alkaline pI. The drawing in Fig. 3is a representation of the distribution of the various spicule matrix proteins combining the results from the three kinds of gels presented in Fig. 2. While the designation of relative abundance is problematic with radiolabeled proteins, it is apparent from the three gels shown in Fig. 2that there are 12 spicule matrix proteins that are more highly radiolabeled by [S]methionine than the rest. These proteins are designated in Fig. 3with stars. Given that the most highly radiolabeled spicule matrix proteins have molecular weights coincident with the most highly staining bands of the one-dimensional SDS-PAGE patterns of spicule matrix proteins shown in Fig. 1B and in Benson et al.(16) , and Venkatesan and Simpson(17) , this suggests that the highly radiolabeled spicule matrix proteins may actually be some of the more prominent spicule matrix proteins.

Figure 3: Diagram of the S. purpuratus spicule matrix proteins. A diagram was drawn summarizing the distribution of the various spicule matrix proteins separated in gels presented in Fig. 2, A-C. Asterisk (*) indicates strongly radiolabeled spicule matrix protein; I, II, and III labels indicate the three spicule matrix proteins that were used to align the proteins that were common among the gels in Fig. 2, A-C. The identity SM30-A, SM30-B, and SM50 spicule matrix proteins are also designated. NEPHGE, non-equilibrium pH gradient gel electrophoresis.

Fig. 2, A-C, also shows approximately three dozen spicule matrix proteins that radiolabel less prominently and they are also represented in Fig. 3. We believe that these less highly radiolabeled proteins are not contaminants since we get very similar patterns on two-dimensional gels with many different batches of radiolabeled spicule matrix proteins. The only differences between batches that we see regularly are slight variations in the relative signals among the various proteins. There is other evidence that the radiolabeled proteins we have identified are not contaminants of non-spicule matrix proteins. Benson et al.(16) using scanning electron microscopy saw that bleach-treated spicule are completely free of embryonic and blastocoelic contaminants. In addition, we found that the two-dimensional pattern of radiolabeled spicule matrix proteins isolated from spicules that are extensively treated with proteinase K subsequent to washing with bleach and prior to demineralization appear identical to radiolabeled proteins isolated from untreated spicules (data not shown).

Identification of SM50 and SM30 Spicule Matrix Proteins

Fig. 4A is a Western blot of a one-dimensional SDS-PAGE separating untreated spicule matrix protein and endoglycosidase F/N-glycosidase F-treated spicule matrix protein. This blot was reacted with the anti-SM50 antiserum, and there is a prominent band of about 48 kDa that is detected with this antiserum. This apparent molecular mass is close to the deduced molecular mass of the mature processed protein encoded by the cloned SM50 cDNA(21) . This observation is the same result that Richardson et al.(29) reported. Western blots using the preimmune serum resulted in no staining (data not shown). It is also apparent from the results in Fig. 4A that the molecular mass does not shift after treatment with endoglycosidase F/N-glycosidase F, indicating that SM50 is not N-glycosylated. Since there is no consensus N-glycosylation site in the deduced SM50 amino acid sequence, this result is expected. There is an additional band visible, other than SM50, in each lane at about 69 kDa in Fig. 4A. This band is not present in other blots using this same antiserum with different spicule matrix preps (e.g.Fig. 6B and Fig. 7B). Also, given the apparent molecular mass of this band, we postulate that this antiserum is reacting with human keratin, a cross-reaction often seen with rabbit polyclonal antisera.

Figure 4: One-dimensional Western blots of spicule matrix proteins using the anti-SM50 and anti-SM30 antisera. A, 0.5 µg of spicule matrix protein and 0.5 µg of endoglycosidase F/N-glycosidase F-treated spicule matrix protein were separated on a 10% SDS-PAGE gel. The gel was then subjected to Western blotting and reacted with the anti-SM50 antiserum. B, 0.5 µg of spicule matrix protein, 0.5 µg of endoglycosidase F/N-glycosidase F-treated spicule matrix protein, and 0.5 µg of alkali-treated spicule matrix protein were separated on a 10% SDS-PAGE gel. The gel was then subjected to Western blotting and was reacted with the anti-SM30 antiserum. The blots in A and B both used alkaline phosphatase-conjugated secondary antibody and BCIP/NBT as substrate to visualize the immunoreactive proteins.

Figure 7: Comparison of embryonic spicule matrix proteins and adult spine integral matrix proteins. A, embryonic spicule matrix protein and adult spine matrix protein were separated on a 10% SDS-PAGE gel and subjected to Western blotting analysis using a polyclonal antiserum raised against all of the spicule matrix proteins. B, embryonic spicule matrix protein and adult spine matrix protein were separated on a 10% SDS-PAGE gel and subjected to Western blotting analysis using the anti-SM50 antiserum. C, embryonic spicule matrix protein and adult spine matrix protein were separated on a 10% SDS-PAGE gel and subjected to Western blotting analysis using the anti-SM30 antiserum. D, endoglycosidase F/N-glycosidase F-treated embryonic spicule matrix protein and adult spine matrix protein were separated on a 10% SDS-PAGE gel and subjected to Western blotting analysis using the anti-SM30 antiserum. Alkaline phosphatase-conjugated secondary antibody and BCIP/NBT as substrate were used to visualize the immunoreactive proteins in these four blots.

Fig. 4B is a one-dimensional Western blot using the anti-SM30 antiserum. Akasaka et al.(23) reported that there are between 2 and 4 copies of SM30 genes per haploid genome and that at least two different SM30 genes, designated SM30- and SM30-, are arranged tandemly in the genome. It was therefore not surprising that a doublet with approximate apparent molecular masses of 43 and 46 kDa reacts with this anti-SM30 antiserum. It also appears that the molecular mass of each doublet member decreases approximately 3-4 kDa after treatment with endoglycosidase F/N-glycosidase F indicating that the two SM30 proteins are N-glycosylated. This glycosidase treatment result is also expected since the deduced amino acid sequence of the SM30- gene and the SM30 pNG7 cDNA each contain one consensus N-glycosylation site(22, 23) . Western blotting using the preimmune serum for the anti-SM30 antiserum resulted in no immunostaining at all (data not shown).

The apparent molecular mass of these anti-SM30 reactive proteins, however, is significantly higher than was expected based on the molecular mass of 30.6 kDa from the SM30 cDNA pNG7 deduced amino acid sequence(22) . The difference in the observed apparent molecular mass and the derived molecular mass cannot be explained solely by N-glycosylation since the removal of N-linked carbohydrate moieties only shifts the molecular mass of the SM30 doublet 3-4 kDa. To determine if O-glycosylation of the SM30 proteins is occurring and contributing to the larger than expected molecular mass, we treated spicule matrix proteins under alkali condition at 37 °C overnight. These conditions should remove any O-glycosylations(33, 34) . The third lane of the Western blot in Fig. 4B shows anti-SM30 antiserum reacted with alkaline-treated spicule matrix proteins. There is no visible alteration of the apparent molecular mass of the SM30 proteins. These results indicate that these anti-SM30 reactive proteins are not O-glycosylated.

Since glycosylation is probably not the reason for the anomalous apparent molecular weight of the SM30 proteins, we devised an experiment to help us determine whether: 1) the SM30 proteins are migrating with a larger apparent molecular mass because of the inherent chemistry of these proteins, 2) the SM30 proteins are being modified further by post-translational modifications other than glycosylation, or 3) the antibody is reacting to other larger spicule matrix proteins that share epitopes with the protein encoded by the SM30 cDNA. To address these issues, capped RNA transcripts were synthesized in vitro using the pNG7 cDNA as template. The pNG7 cDNA encodes a full-length SM30 protein(22) . This SM30 RNA and [S]methionine were then co-microinjected into Xenopus oocytes. Radiolabeled proteins secreted by SM30 RNA injected and control oocytes were collected, and the secreted proteins were then subjected to immunoprecipitation using the anti-SM30 antiserum. The immunoprecipitates were then separated by SDS-PAGE. The results of this experiment are shown in Fig. 5A. From this gel, it is apparent that a 46-kDa secreted protein is synthesized by the oocytes injected with pNG7 SM30 mRNA. Fig. 5A also shows that the SM30 protein synthesized by the Xenopus oocyte decreases in apparent molecular mass by 3-4 kDa when it is treated with endoglycosidase F/N-glycosidase F.

Figure 5: In vitro translation of SM30 RNA by Xenopus oocytes and reticulocyte lysate. SM30 RNA was synthesized in vitro using the SM30 cDNA clone pNG7 (22) as the template. A, Xenopus oocytes were microinjected with in vitro synthesized SM30 RNA and [S]methionine. The supernatant of the SM30 RNA injected oocytes which contained radiolabeled, secreted proteins were collected 2 days later (control). A portion of this supernantant was treated with endoglycosidase F/N-glycosidase F (glycosidase treated). The untreated and glycosidase-treated supernatants were then subjected to immunoprecipitation using the preimmune and anti-SM30 antisera. The resulting precipitates were separated on a 10% SDS-PAGE gel, processed for fluorography, and exposed to x-ray film. The anti-SM30 antiserum specifically immunoprecipitates a broad 46-kDa band from the untreated oocyte supernantant (indicated as untreated SM30). This band decreases in apparent molecular mass approximately 3-4 kDa when the oocyte supernantant is glycosidase treated (indicated as treated SM30). B, rabbit reticulocyte lysate was used to translate in vitro SM30 RNA, control RNA (Xenopus elongation factor 1), and no added RNA in the presence of [S]methionine. Equal portions of the lysates that contained the various radiolabeled protein products were then separated on a 10% SDS-PAGE gel and processed for fluorography. The protein product synthesized using the SM30 RNA has an approximately 32-kDa apparent molecular mass.

SM30 RNA synthesized from the pNG7 cDNA was also translated in vitro using rabbit reticulocyte lysate, which does not post-translationally modify translation products. Fig. 5B shows that the reticulocyte lysate synthesizes an approximately 32-kDa product which is the molecular masss of the deduced amino acid sequence of the pNG7 SM30 cDNA that contains its signal sequence. Given these results, it appears that one or more post-translation modifications other than N-glycosylation produces the anomalous electrophoretic migration of the SM30 proteins. Since the Xenopus oocyte seems to be modifying the pNG7 SM30 protein correctly, a convenient in vitro assay exists to determine the nature of the post-translational modifications occurring to SM30 proteins. We are presently attempting to identify the as of yet unknown post-translational modification(s) of the SM30 proteins with this system.

To further determine if the observed SM30 doublet was real or apparent, we also ran gels with spicule matrix proteins subjected to higher amounts of reducing agent and higher SDS concentrations. However, these treatments did not alter the migration or presence of the anti-SM30 reactive doublet (data not shown). Since the 43- and 46-kDa anti-SM30 reactive proteins appear to encode genuine SM30 proteins, these proteins are designated SM30-A and SM30-B, respectively.

To determine which of the spots on the two-dimensional gels of the spicule matrix proteins are actually the SM30 and SM50 proteins, Western blots of two-dimensional gels were done. Fig. 6A is a two-dimensional Western blot using the anti-SM30 antiserum. The first dimension of this gel was an equilibrium isoelectric focusing gel using ampholytes with a range of 2.5-6.5 (similar to the first dimension of the gel in Fig. 2B). From this blot it is apparent that the anti-SM30 antiserum reacts with two closely migrating spots at the acidic end of the gel with the same apparent molecular weight as the SM30-A and SM30-B proteins in Fig. 4B. When this gel is compared to a gel separating labeled spicule matrix proteins that was run in parallel, it is apparent that SM30-A and SM30-B are two of the more acidic and highly radiolabeled spicule matrix proteins. The identity of the SM30-A and SM30-B proteins are marked in Fig. 3.

Fig. 6B is a two-dimensional Western blot using the anti-SM50 antiserum. The first dimension of this blot was a non-equilibrium pH gradient gel using ampholytes with a pH range of 3-10 and the gel run in the basic direction (similar to the first dimension of the gel in Fig. 2C). From this blot it is apparent that the anti-SM50 antiserum reacts with a single protein that migrates with an apparent molecular mass of 48 kDa at the most alkaline portion of the gel. When compared to a gel separating labeled spicule matrix protein that was run in parallel, it is apparent that SM50 is the spicule matrix protein with the most alkaline isoelectric point. This observation is consistent with the pI of the deduced amino acid sequence of the SM50 cDNA being approximately 12(21) . The identity of the SM50 protein is marked in Fig. 3.

Comparison of Spicule Matrix Proteins with the Integral Matrix Proteins of Adult Spines

Comparison of one-dimensional Western blots of embryonic spicule matrix protein and adult spine integral matrix protein, using an antiserum raised against all of the sea urchin embryonic spicule matrix proteins, is shown in Fig. 7A. This figure reveals extensive cross-reactivity between the two tissues. Qualitatively, the most immunoreactive proteins in adult spine matrix seem to have apparent molecular masses larger than those from the spicule matrix. One exception is a band at approximately 120 kDa which has a similar sized counterpart in the spicule matrix protein. We know, however, that the embryonic spicule matrix band at 120 kDa is composed of a few different proteins of that same size in the spicule matrix (see Fig. 2and Fig. 3).

A Western blot using anti-SM50 antiserum (Fig. 7B) reveals that SM50 has the same apparent molecular mass in spine matrix as it does in spicule matrix, an observation seen by Richardson et al.(29) using a different anti-SM50 antiserum. One-dimensional blot analysis of spicule and spine matrix proteins using the anti-SM30 antiserum reveals differences between these two tissues. This result is shown in Fig. 7, C and D. While anti-SM30 reacts with spicule matrix bands with apparent molecular masses of 43 and 46 kDa, this antiserum reacts with a doublet of 46 and 49 kDa in spine matrix (see Fig. 7C). These two sets of SM30 doublets in Fig. 7C are located at exactly the same apparent molecular mass as two prominent immunoreactive doublets in the spicule and spine lanes in Fig. 7A that reacted with the anti-total spicule matrix antiserum. To determine whether the apparent size difference in the doublets that react with anti-SM30 is caused by differential glycosylation, glycosidase-treated spicule and spine matrix protein were analyzed by Western blotting. As can be seen in Fig. 7D, both spicule and spine doublets have reduced apparent molecular masses of about 3 kDa each after glycosidase treatment. However, a size difference between the SM30 doublets present in spicule and spine matrix persists. This finding suggests that the difference in apparent molecular weight is not because of differential N-glycosylation. Rather, it suggests that the 49-kDa protein is a different form of SM30 and/or it is post-translationally modified (other than N-glycosylation) differently than SM30-A or SM30-B. We designate this apparent adult isoform of SM30 as SM30-C. The results in Fig. 7, C and D, also suggest that the 43-kDa SM30-A protein is an embryonic isoform of SM30 since there is no protein of that size in the adult spine matrix lane. The 46-kDa protein, SM30-B, is apparently expressed in both spicules and adult spines. Of course, at this point, we cannot rule out the possibility that the 46-kDa proteins in the embryonic spicule matrix and the adult spine matrix are encoded by different genes. Akasaka et al.(23) have shown that there are up to four different SM30 genes. But, at this time, we have no reason to invoke that possibility.

Comparison of the Sequence of the SM30 Proteins with C-type Lectin Family of Proteins

Figure 8: The similarity between SM30- protein and proteins containing a C-type lectin carbohydrate recognition domain. A region of SM30- protein is aligned with various proteins containing a C-type lectin carbohydrate recognition domain. Amino acids that are boxed in black indicate identity with the SM30- protein. Amino acids that are boxed in gray indicate a conservative amino acid substitution. Numbers in brackets next to protein names indicate the residue numbers of the various proteins that are aligned in the figure. Dash(-) indicates gaps that were added to help align the sequences. Numbers at the top of the alignment refer to amino acid residue number for the SM30- protein (accession number P28163). Protein names are in Swiss Prot format. LITHRAT, rat lithostathine precursor (accession number P10758); LITHHUMAN, human lithostathine precursor (accession number P05451); MANRHUMAN, human macrophage mannose receptor (accession number P22897); PAP1HUMAN, human pancreatitis-associated protein 1 precursor (accession number Q06141); LITHBOVIN, bovine lithostathine precursor (accession number P23132); PAP1RAT, rat pancreatitis-associated protein 1 precursor (accession number P25031); PAP1MOUSE, mouse pancreatitis-associated protein 1 precursor (accession number P35230); TETNHUMAN, human tetranectin precursor (accession number P05452); LECAPLEWA, Pleuradeles waltii lectin precursor (accession number Q02988); LECEANTCR, Anthocidaris crassispina echinoidin (accession number P06027). Arrows above the SM30- protein sequence indicate residues that are different in the protein encoded by pNG7 SM30 cDNA. The different amino acid residues in the pNG7 protein are indicated on top of the arrows. Symbols for the conserved structure and amino acid residues of C-type lectins determined by Weis et al.(44) : = aliphatic; = aromatic; = aliphatic or aromatic; Z = E or Q; B = D or N; W = side chain containing oxygen (D, N, or Q); W, D, P, and C = single letter symbols of amino acids; 1, first Ca binding site; 2, second Ca site; S, small hydrophobic core; L, large hydrophobic core.

When SM30- was initially cloned and characterized, it was noted there were nine amino acid differences between the protein encoded by it and the protein encoded by the SM30 pNG7 cDNA(23) . It was unclear if these differences reflected the well know polymorphisms of sea urchin genomic sequences or if it reflected that SM30- and pNG7 encode different SM30 proteins. It is interesting to note that seven of the nine amino acid differences between the SM30- protein and the pNG7 SM30 protein occur over the approximately 40% portion of the SM30 proteins that is similar to the CRD of C-type lectins. Arrows in Fig. 8indicate where these differences occur. Most of these differences alter residues shown in the present study to be conserved between SM30- protein and C-type lectins or in residues shown by Weis et al.(44) to be conserved among C-type lectins. While it is premature to conclude too much from these differences, these findings suggest that SM30- and pNG7 SM30 may be different forms of the SM30 protein.

DISCUSSION

Acidic integral matrix proteins isolated from calcitic adult sea urchin exoskeletal tissues have been shown to bind to specific faces of calcite crystal in vitro(4) . It has also been shown that the intercalation of the matrix proteins within the calcite is responsible for some of the physical properties of adult sea urchin skeletal elements(5, 6) . However, these studies did not characterize the integral matrix proteins other than determining partial amino acid composition. There is evidence that the integral matrix proteins of adult mineralized tissues and the integral matrix proteins of embryonic spicules control similar physical properties of these calcitic tissues (6) . In addition, transcripts complementary to the previously cloned spicule matrix SM50 and SM30 cDNAs are expressed in adult mineralized tissues(22, 29) . Therefore we decided to further characterize the embryonic spicule matrix proteins and compare them to adult integral matrix proteins.

Our studies presented here show that there are many more spicule matrix proteins than were detected in previous studies. A dozen prominently radiolabeled spicule matrix proteins and some three dozen less prominently radiolabeled spicule matrix proteins are detected. The apparent molecular mass of the proteins range from 20 kDa to over 100 kDa. Most of these proteins, including the more prominent ones, are acidic. We also show that there are several less intensely radiolabeled spicule matrix proteins that have more alkaline isoelectric points. The predicted chemistry of the spicule matrix proteins is consistent with that found for other integral matrix proteins of calcified structures for other organisms(8, 9) . The results presented in this article further indicate that the adult integral matrix proteins are similar to embryonic spicule matrix proteins since there is significant Western blot cross-reactivity with the anti-total spicule matrix antiserum. However, since the one-dimensional SDS-PAGE Western antibody staining pattern is different from spicule matrix protein's pattern, this suggests that there are several different integral matrix proteins, or at least several different forms of integral matrix protein, in the adult spine.

Western blot analysis of the previously cloned SM50 gene shows that the SM50 protein is the spicule matrix protein with the most alkaline isoelectric point and that it has an apparent molecular mass of approximately 48 kDa. This result is consistent with the observations of Livingston et al.(28) and Richardson et al.(29) . Livingston et al.(28) cloned the L. pictus homologue of SM50. This 34-kDa protein, designated LSM34, is also not glycosylated and has an alkaline pI as determined from its derived amino acid sequence. The glycosidase results of the present paper and Livingston et al.(28) are to be expected since there is no N-linked glycosylation consensus sequence in the derived amino acid sequence of the S. purpuratus SM50 cDNA (21) and the L. pictus LSM34 cDNA(28) . Richardson et al.(29) generated an anti-SM50 specific polyclonal antiserum and they observed that it reacted with a single band of approximately 48 kDa in S. purpuratus embryonic spicules and adult spines.

It is particularly interesting that the Western blotting analysis using the anti-SM30 antiserum reveals that there are multiple forms of the SM30 protein in the sea urchin embryonic spicule matrix and in the adult spine matrix. This result is consistent with our assertion that adult spine integral matrix contains some proteins that are different from embryonic spicule matrix. When the SM30 cDNA was originally isolated, Northern blots revealed only one band hybridizing with the SM30 cDNA probe(22) . However, recent work (23) has demonstrated that SM30 is a small multigene family with between two and four copies of the SM30 gene present per haploid genome of S. purpuratus. A genomic clone was also isolated and it was found that at least two SM30 genes are tandemly arranged (designated SM30- and SM30-). Therefore, it is consistent, given these findings, to see a doublet of SM30 proteins in the spicule matrix of 43 and 46 kDa (designated as SM30-A and SM30-B, respectively) and a doublet of SM30 protein in adult spine matrix of 46 and 49 kDa (designated SM30-B and SM30-C, respectively).

Harkey et al.(27) reported that SM50 and the PM27 proteins are similar to C-type lectin proteins and we now report that SM30 is also similar to the CRD domain of C-type lectins. The C-type lectins are a very heterogenous family of proteins that are most often involved in recognition events outside cells (see for reviews, (45, 46, 47) ). Usually these recognition events are mediated through a Ca dependent binding to carbohydrate moieties. However, certain C-type lectins have also been shown to bind proteins directly (48) , and to act as antifreeze molecules in coldwater fish(49) .

It is interesting to note that two types of C-type lectins that were found to be most similar to SM30- protein in our studies here have previously been found to be involved in the formation of mineralized structures within animals. One of these C-type lectins is tetranectin which is a blood and extracellular matrix component. Wewer et al.(50) reported localization of tetranectin in mineralizing mouse bone osteoblastic cells that were differentiating in vitro. However, the precise molecular role tetranectin plays in osteogenesis remains unknown. The other C-type lectin previously found to be involved in mineralized structures is called lithostathine (formerly called pancreatic stone protein(51) ). Lithostathine is present in pancreatic juices and it is believed to bind calcium carbonate and prevent calcium carbonate from precipitating and forming calcitic pancreatic stones (see for review, (52) ). However, the 11 amino acid peptide sequence at the amino end of lithostathine that is thought to interact with calcium carbonate is not within the CRD of lithostathine and it is not well conserved in SM30 or SM50 (data not shown). So whether there is homology of function of SM30- protein and lithostathine to go along with the similarity of their sequences remains enigmatic. These findings, however, raise several interesting questions. Could there be other regions of these spicule matrix proteins that bind calcium carbonate? Could the similarity to C-type lectins be reflective of the SM30 and SM50 proteins binding the carbohydrate moieties of other spicule matrix proteins? SM30- protein has imperfect matching at the calcium ion binding sites typically found in C-type lectins. Does SM30- protein bind Ca? Many attempts in our laboratory over the years by different techniques have failed to reveal Ca binding to SM30. Why are so many of the known proteins which are closely associated with spicule formation similar to C-type lectins? Experiments are underway to address these questions.

It remains undetermined which of the SM30 proteins identified in the present paper are encoded by the SM30-, SM30-, or the pNG7 SM30 cDNA. Results presented in this paper point out that 7 out of 9 differences in the amino acid sequence between the protein encoded by the pNG7 SM30 cDNA and the protein encoded by SM30- occur over the approximately 40% portion of the SM30 proteins that comprise the region similar to the CRD of C-type lectins. Most of these differences alter conserved residues suggesting that they may encode for different forms of the SM30 proteins. In addition to the studies presented here, unpublished RNase protection studies show that SM30- gene transcript is expressed in the sea urchin embryo and not in the adult spine. ()Since SM30-A protein seems to be present in embryonic spicule and not in adult spines, this unpublished finding suggests that SM30-A may be encoded by the SM30- genomic sequence. The complete sequence of SM30- is not yet known (23) and results presented in the current paper suggest that there is at least a third SM30 gene that has not been cloned. Therefore, until the complete sequence of all of the SM30 genomic sequences as well as their expression patterns are known, we cannot be sure if SM30- encodes SM30-A or if pNG7 SM30 cDNA encodes for a SM30 protein different from the one encoded by SM30-. Studies addressing these issues are underway.

Our enumeration, characterization, and comparisons of spine and spicule matrix proteins complement studies of Berman et al.(4, 5, 6) . They observed that sea urchin integral matrix proteins are not as acidic as some other invertebrate integral matrix proteins. We have found that there is a range of acidic to alkaline pI values for the spicule matrix proteins, although the net pI of the proteins as a whole is acidic(16) . Berman et al.(4) showed that the integral matrix proteins from adult sea urchin mineralized tissues, but not the more acidic integral proteins from the mollusc Mytilus californus prismatic layer, were able to bind specific calcite crystal faces, even though the mollusc proteins are known nucleators of calcite when they are adsorbed on a rigid substrate. Berman et al.(6) also studied differences in texture of calcite crystals among calcitic tissues from different taxonomic groups, including analysis of sea urchin embryo spicules and adult spines. They showed that the manipulation of crystal structure is under biological control and that the integral matrix proteins probably play a role in this control. The studies presented here provide a fuller characterization of the embryo spicule and adult spine matrix proteins integral matrix proteins. Future studies looking at physical interactions of individual spicule matrix proteins with calcite crystals in vitro will be particularly informative.

FOOTNOTES

*

This work was supported by National Institutes of Health Grant HD 15043 (to F. H. W.) and National Aeronautics and Space Administration Grant NAG 5-72 (to F. H. W.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore by hereby marked ``advertisement'' in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

§

To whom reprint requests and correspondence should be sent: University of California, Berkeley, Dept. Molecular & Cell Biology, Life Sciences Addition, Rm. 379, Berkeley, CA 94720. Tel.: 510-642-2927; Fax: 510-643-6791; ckillian{at}uclink2.berkeley.edu.

()

The abbreviations used are: PAGE, polyacrylamide gel electrophoresis; CRD, carbohydrate recognition domain; BCIP/NBT, 5-bromo-4-chloro-3-indolyl phosphate/nitro blue tetrazolium chloride.

()

C. E. Killian and F. H. Wilt, unpublished data.

ACKNOWLEDGEMENTS

REFERENCES

1. Emlet, R. B. (1982) Biol. Bull. 163, 264-275 [Abstract/Free Full Text]
2. Okazaki, K., and Inoué, S. (1976) Dev. Growth and Differ. 18, 413-434
3. Okazaki, K., McDonald, K., and Inou é , S. (1980) in The Mechanisms of Biomineralization in Animals and Plants (Omori, M., and Watabe, N., eds) pp. 159-168, Tokai University Press, Tokyo
4. Berman, A., Addadi, L., and Weiner, S. (1988) Nature 331, 546-548 [CrossRef]
5. Berman, A., Addadi, L., Kvick, Å., Leiserowitz, L., Nelson, M., and Weiner, S. (1990) Science 250, 664-667 [Abstract/Free Full Text]
6. Berman, A., Hanson, J., Leiserowitz, L., Koetzle, T., Weiner, S., and Addadi, L. (1993) Science 259, 776-779 [Abstract/Free Full Text]
7. Aizenberg, J., Hanson, J., Ilan, M., Leiserowitz, L., Koetzle, T. F., Addadi, L., and Weiner, S. (1995) Faseb J. 9, 262-268 [Abstract]
8. Weiner, S. (1984) Am. Zool. 24, 945-951
9. Weiner, S. (1986) CRC Crit. Rev. Biochem. 20, 365-408 [Medline] [Order article via Infotrieve]
10. Okazaki, K. (1975) Am. Zool. 15, 567-581
11. Decker, G., and Lennarz, W. J. (1988) Development 103, 231-247 [Medline] [Order article via Infotrieve]
12. Ettensohn, C. A. (1991) in Cell-Cell Interactions in Early Development, 49th Symposium of the Society for Developmental Biology (Gerhart, J., ed) pp. 175-201, Wiley-Liss, Inc., New York
13. Benson, S. C., and Wilt, F. H. (1992) in Calcification in Biological Systems (Bonucci, E., ed) pp. 157-178, CRC Press, Ann Arbor, MI
14. Wilt, F. H., Killian, C. E., and Livingston, B. T. (1993) in Chemistry and Biology of Mineralized Tissues (Slavkin, H., and Price, P., eds) pp. 85-92, Excerpta Medica, New York
15. Benson, S. C., Jones, E. M., Benson, N. C., and Wilt, F. (1983) Exp. Cell Res. 148, 249-253 [CrossRef][Medline] [Order article via Infotrieve]
16. Benson, S. C., Benson, N. C., and Wilt, F. H. (1986) J. Cell Biol. 102, 1878-1886 [Abstract/Free Full Text]
17. Venkatesan, M., and Simpson, R. T. (1986) Exp. Cell Res. 166, 259-264 [CrossRef][Medline] [Order article via Infotrieve]
18. Sucov, H. M., Benson, S., Robinson, J. J., Britten, R. J., Wilt, F., and Davidson, E. H. (1987) Dev. Biol. 120, 507-519 [CrossRef][Medline] [Order article via Infotrieve]
19. Benson, S., Sucov, H., Stephens, L., Davidson, E., and Wilt, F. (1987) Dev. Biol. 120, 499-506 [CrossRef][Medline] [Order article via Infotrieve]
20. Killian, C. E., and Wilt, F. H. (1989) Dev. Biol. 133, 148-156 [CrossRef][Medline] [Order article via Infotrieve]
21. Katoh-Fukui, Y., Noce, T., Ueda, T., Fujiwara, Y., Hashimoto, N., Higashinakagawa, T., Killian, C., Livingston, B., Wilt, F., Benson, S., Sucov, H., and Davidson, E. H. (1991) Dev. Biol. 145, 201-202 [CrossRef][Medline] [Order article via Infotrieve]
22. George, N. C., Killian, C. E., and Wilt, F. H. (1991) Dev. Biol. 147, 334-342 [CrossRef][Medline] [Order article via Infotrieve]
23. Akasaka, K., Frudakis, T. N., Killian, C. E., George, N. C., Yamasu, K., Khaner, O., and Wilt, F. (1994) J. Biol. Chem. 269, 20592-20598 [Abstract/Free Full Text]
24. Sucov, H. M., Hough-Evans, B. R., Franks, R. R., Britten, R. J., and Davidson, E. H. (1988) Genes and Develop. 2, 1238-1250 [Abstract/Free Full Text]
25. Makabe, K. W., Kirchhamer, C. V., Britten, R. J., and Davidson, E. H. (1995) Development 121, 1957-1970 [Abstract]
26. Frudakis, T. N., and Wilt, F. H. (1995) Dev. Biol. 172, 230-241 [CrossRef][Medline] [Order article via Infotrieve]
27. Harkey, M. A., Klueg, K., Sheppard, P., and Raff, R. A. (1995) Dev. Biol. 168, 549-566 [CrossRef][Medline] [Order article via Infotrieve]
28. Livingston, B. T., Shaw, R., Bailey, A., and Wilt, F. (1991) Dev. Biol. 148, 473-480 [CrossRef][Medline] [Order article via Infotrieve]
29. Richardson, W., Kitajima, T., Wilt, F., and Benson, S. (1989) Dev. Biol. 132, 266-269 [CrossRef][Medline] [Order article via Infotrieve]
30. Benson, S., Smith, L., Wilt, F., and Shaw, R. (1990) Exp. Cell Res. 188, 141-146 [CrossRef][Medline] [Order article via Infotrieve]
31. Meier, T., Arni, S., Malarkannan, S., Poincelet, M., and Hoessli, D. (1992) Anal. Biochem. 204, 220-226 [CrossRef][Medline] [Order article via Infotrieve]
32. Nesbitt, S., and Horton, M. A. (1992) Anal. Biochem. 206, 267-272 [CrossRef][Medline] [Order article via Infotrieve]
33. Florman, H. M., and Wassarman, P. M. (1985) Cell 41, 313-324 [CrossRef][Medline] [Order article via Infotrieve]
34. Sharon, N. (1975) Complex Carbohydrates, Their Chemistry, Biosynthesis, and Function, pp. 73-78, Addison-Wesley Publishing Co., Reading, MA
35. O'Farrell, P. H. (1975) J. Biol. Chem. 250, 4007-4021 [Abstract/Free Full Text]
36. Laemmli, U. K. (1970) Nature 227, 680-685 [CrossRef][Medline] [Order article via Infotrieve]
37. Dreyfus, G., Adams, S. A., and Choi, Y. D. (1984) Mol. Cell. Biol. 4, 415-423 [Abstract/Free Full Text]
38. Laskey, R. A., and Mills, A. D. (1975) Eur. J. Biochem. 56, 335-341 [Medline] [Order article via Infotrieve]
39. Harlow, E., and Lane, D. (1988) Antibodies: A Laboratory Manual, Cold Spring Harbor Laboratory, Cold Sping Harbor, NY
40. Towbin, H., Stachelin, T., and Gordon, J. (1979) Proc. Natl. Acad. Sci. U. S. A. 76, 4350-4354 [Abstract/Free Full Text]
41. Dunbar, B. S., Kimura, H., and Timmons, T. M. (1990) Methods Enzymol. 182, 441-459 [Medline] [Order article via Infotrieve]
42. Raup, D. M. (1966) in Physiology of Echinodermata (Boolootian, R. A., ed) pp. 379-395, Intersciences Publishers, London
43. Harkey, M. A., and Whiteley, A. H. (1983) Dev. Biol. 100, 12-28 [CrossRef][Medline] [Order article via Infotrieve]
44. Weis, W. I., Kahn, R., Fourme, R., Drickamer, K., and Hendrickson, W. A. (1991) Science 254, 1608-1615 [Abstract/Free Full Text]
45. Drickamer, K. (1993) Prog. Nucleic Acids Res. Mol. Biol. 45, 207-232 [Medline] [Order article via Infotrieve]
46. Drickamer, K. (1993) Curr. Opin. Struct. Biol. 3, 393-400
47. Drickamer, K., and Taylor, M. E. (1993) Ann. Rev. Cell Biol. 9, 237-264 [CrossRef]
48. Bettler, B., Texido, G., Raggini, S., Rüegg, D., and Hofstetter, H. (1992) J. Biol. Chem. 267, 185-191 [Abstract/Free Full Text]
49. Ng, N. F. L., and Hew, C. L. (1992) J. Biol. Chem. 267, 16069-16075 [Abstract/Free Full Text]
50. Wewer, U. M., Ibaraki, K., Schjørring, P., Durkin, M. E., Young, M. F., and Albrechtson, R. (1994) J. Cell Biol. 127, 1767-1775 [Abstract/Free Full Text]
51. Sarles, H., Dagorn, J. C., Giorgi, D., and Bernard, J. P. (1990) Gastroentroenterology 99, 900-901
52. Dagorn, J. C. (1993) in The Pancreas (Go, V. L. W., DiMagno, E. P., Gardner, J. D., Lebenthal, E., Reber, H. A., and Scheele, G. A., eds) pp. 253-263, Raven Press, Ltd., New York

CiteULike

Complore

Connotea

Del.icio.us

Digg

Reddit

Technorati

This article has been cited by other articles:

				Y. Politi, R. A. Metzler, M. Abrecht, B. Gilbert, F. H. Wilt, I. Sagi, L. Addadi, S. Weiner, and P. U. P. A. Gilbert Transformation mechanism of amorphous calcium carbonate into calcite in the sea urchin larval spicule PNAS, November 11, 2008; 105(45): 17362 - 17366. [Abstract] [Full Text] [PDF]

				M. Yajima and M. Kiyomoto Study of larval and adult skeletogenic cells in developing sea urchin larvae. Biol. Bull., October 1, 2006; 211(2): 183 - 192. [Abstract] [Full Text] [PDF]

				X. Li, D. Zhang, V. J. Lynch-Holm, T. W. Okita, and V. R. Franceschi Isolation of a Crystal Matrix Protein Associated with Calcium Oxalate Precipitation in Vacuoles of Specialized Cells Plant Physiology, October 1, 2003; 133(2): 549 - 559. [Abstract] [Full Text] [PDF]

				Y. Dauphin Soluble Organic Matrices of the Calcitic Prismatic Shell Layers of Two Pteriomorphid Bivalves. PINNA NOBILIS AND PINCTADA MARGARITIFERA J. Biol. Chem., April 18, 2003; 278(17): 15168 - 15177. [Abstract] [Full Text] [PDF]

				R. Lakshminarayanan, S. Valiyaveettil, V. S. Rao, and R. M. Kini Purification, Characterization, and in Vitro Mineralization Studies of a Novel Goose Eggshell Matrix Protein, Ansocalcin J. Biol. Chem., January 24, 2003; 278(5): 2928 - 2936. [Abstract] [Full Text] [PDF]

				T. Nagai, S.-i. Kawabata, F. Shishikura, and H. Sugita Purification, Characterization, and Amino Acid Sequence of an Embryonic Lectin in Perivitelline Fluid of the Horseshoe Crab J. Biol. Chem., December 31, 1999; 274(53): 37673 - 37678. [Abstract] [Full Text] [PDF]

				L. Ameye, R. Hermann, C. Killian, F. Wilt, and P. Dubois Ultrastructural Localization of Proteins Involved in Sea Urchin Biomineralization J. Histochem. Cytochem., September 1, 1999; 47(9): 1189 - 1200. [Abstract] [Full Text]

This Article Abstract Full Text (PDF) Submit a Letter to Editor Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Permissions Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Killian, C. E. Articles by Wilt, F. H. Search for Related Content PubMed PubMed Citation Articles by Killian, C. E. Articles by Wilt, F. H. Social Bookmarking                      What's this?